| 16:07 | kaipee_mobile | I seem to be having filesystem corruption, which i have been advised may be caused by hard powering-off the system |
| 16:07 | curlymo | what distro? |
| 16:07 | kaipee_mobile | Is there anything i can do to overcome this? |
| 16:07 | kaipee_mobile | curlymo openelec |
| 16:08 | curlymo | you could try XBian, we use btrfs as the root filesystem. I never got corruption since. |
| 16:08 | jnettlet | use reboot -p then power off the device |
| 16:08 | curlymo | i'm also too lazy to do that :p |
| 16:08 | jnettlet | I assume the data corruption is on your data partition |
| 16:08 | jnettlet | it must be because openelec uses a ramdisk |
| 16:09 | fritsch | ... |
| 16:10 | fritsch | hard power off any mounted journaling fs while doing something that reads and writes you will get those errors |
| 16:10 | curlymo | not if the filesystem can check if those reads/writes succeeded or failed with checksums |
| 16:10 | fritsch | if you are lucky the journal can repair them |
| 16:11 | curlymo | and ofc copy-on-write |
| 16:11 | kaipee_mobile | [ 115.007009] EXT4-fs error (device mmcblk0p2): ext4_lookup:1429: inode #523286: comm connmand: deleted inode referenced: 523292 |
| 16:11 | kaipee_mobile | curlymo dmesg is pretty much flooded with those |
| 16:13 | curlymo | indeed possible corruption? Can you fsck the sdcard in a linux machine? |
| 16:13 | kaipee_mobile | curlymo yeah i have a few Arch boxes |
| 16:14 | kaipee_mobile | I'm mite concerned with preventing future corruption |
| 16:14 | kaipee_mobile | *more |
| 16:14 | curlymo | gracefull shutdowns or switch FS (like XBian did) |
| 16:14 | fritsch | curlymo: any performance impact btrfs has? |
| 16:15 | kaipee_mobile | Can i configure OE to use a custom command for the Power button (on UI)? |
| 16:16 | kaipee_mobile | I can only presume this is a common issue with the cubox because it is an 'always on' device |
| 16:16 | kaipee_mobile | ...except that nobody really notices |
| 16:17 | curlymo | @fritsch, no |
| 16:17 | fritsch | kaipee_mobile: just press poweroff |
| 16:17 | fritsch | it will halt |
| 16:17 | fritsch | when it's frozen, power it off |
| 16:17 | fritsch | e.g. remove the plug |
| 16:18 | kaipee_mobile | fritsch grand, will do. I'll need to educate the housemates about switching off the mains power supply all the time |
| 16:18 | fritsch | you can also give btrfs a try |
| 16:18 | fritsch | if you want |
| 16:18 | curlymo | in OE? |
| 16:18 | fritsch | would be good to know if that really compensates your needs |
| 16:19 | fritsch | nope, he can use xbian if he likes to |
| 16:19 | fritsch | it open source :-) choose what fits |
| 16:19 | kaipee_mobile | Well i have another sdcard, i might put Xbian on it for fun |
| 16:19 | fritsch | yeah, give that a try |
| 16:19 | jnettlet | f2fs is actually the recommended filesystem for SDHC media. |
| 16:20 | jnettlet | Samsung has done a good job, except everyone is shying away from it for stupid reasons |
| 16:20 | curlymo | OSMC tried it but got back from it |
| 16:21 | fritsch | blkid does not recognize it |
| 16:21 | jnettlet | I have been using it on a bunch of my Android devices, and my main CBi device |
| 16:22 | jnettlet | I think the big drawback is you can't resize it yet. |
| 16:22 | fritsch | curlymo: why? |
| 16:22 | fritsch | jnettlet: yeah, that's a bummer for an automated installation, e.g. resize on first boot |
| 16:23 | curlymo | don't know |
| 16:23 | curlymo | just heard |
| 16:23 | fritsch | alpha1 too unstable cause of f2fs :-) |
| 16:24 | jnettlet | fritsch, easier to just create the data partition on boot and make a clean filesystem. Bad unix sysadmin skills mixing binaries and data that require resizing FS's all the time. |
| 16:24 | fritsch | jnettlet: thought of that while writing |
| 16:25 | fritsch | i think we need some real tests. e.g. take ext4, copy something over network to that partition while reading something else |
| 16:25 | fritsch | and pull the plug |
| 16:25 | curlymo | @jnettlet, depends. In BTRFS you can just have subvols like in zfs. Live resizing is absolutely no issue there. |
| 16:25 | fritsch | do it 10 times and see what happens |
| 16:28 | curlymo | nothing :) |
| 16:28 | fritsch | curlymo: nothing for both or total failure for ext4? :-) |
| 16:30 | Catwich | on btrfs: I am not happy! I have a completely dead btrfs partition right now in this computer |
| 16:30 | Catwich | and it is teh second time it happened to me |
| 16:30 | Catwich | so I stay away from it now |
| 16:30 | curlymo | what version? |
| 16:30 | Catwich | I realized I couldn't recover data from the home subvolume |
| 16:30 | Catwich | so I lost everything important |
| 16:30 | curlymo | @fritsch, also running a HB i2eX as offsite ZFS backup :D |
| 16:30 | Catwich | it was kernel 3.16.6 |
| 16:31 | Catwich | not sure of anything more specific |
| 16:31 | kaipee_mobile | Right I'm off, cheers for the tips guys |
| 16:32 | fritsch | let me find that kernel mail about btrfs |
| 16:32 | fritsch | they weren't happy either |
| 16:33 | fritsch | curlymo: https://bbs.archlinux.org/viewtopic.php?id=182505 |
| 16:33 | fritsch | curlymo: i think i don't want to throw that on the users ... |
| 16:33 | kaipee_mobile | I'll try reflashing OE later, also I'll try Xbian ? |
| 16:33 | jnettlet | lol. I think that has been my mantra for the last two years |
| 16:34 | curlymo | we never had issues |
| 16:34 | curlymo | using it over a year now |
| 16:34 | fritsch | curlymo: http://marc.info/?l=linux-btrfs&m=142015947409889&w=2 <- does that load? |
| 16:35 | jnettlet | not for me |
| 16:36 | curlymo | will check later |
| 16:36 | curlymo | currently checking the radxa rock |
| 16:36 | fritsch | ext4 has been proven stable over many years - the more i google: btrfs recovery |
| 16:36 | fritsch | :-) |
| 16:36 | fritsch | the more afraid I get |
| 16:37 | curlymo | i can only speek from my experience |
| 16:37 | curlymo | rather use zfs, but that's not quite an option performance wise |
| 16:37 | fritsch | http://www.mail-archive.com/[email protected]/msg37924.html <- also interesting |
| 16:38 | curlymo | when you are experienced with zfs, btrfs is hell |
| 16:38 | curlymo | but it solves the issues we had |
| 18:27 | vpeter | Are there any pictures for disassembled cubox-i? |
| 18:27 | fritsch | vpeter: you could make some :-) |
| 18:30 | vpeter | haha |
| 18:30 | vpeter | Probably I'm not the first one with this idea. |
| 18:33 | vpeter | It should look something like this: http://www.cnx-software.com/wp-content/uploads/2014/08/STACK_Box_Disassembled_Large-e1408510607847.jpg |
| 18:40 | palatin | Hi, does anyone know of a working kernel config file for recent kernel releases on cubox v1 ? Headless is fine |
| 18:55 | jnettlet | palatin, somewhere around 3.17 upstream broke I believe. We haven't had anytime to look into it. |
| 18:55 | jnettlet | I would checkout the 3.16.y branch from linux-stable |
| 19:08 | palatin | ok, I have tried to update from 3.6 to vanilla 3.12 and 3.14 a few times to no avail, The kernel does not seem to load I don't have any clue |
| 19:08 | palatin | I gather there is something wrong with my config |
| 19:09 | jnettlet | palatin, I think rmk just got a new production Cubox v1 so anything that is broken may be fixed relatively soon. |
| 19:10 | jnettlet | I will ping him about it for you. |
| 19:10 | palatin | thanks |
| 20:37 | soppman | hello |
| 20:37 | soppman | anyone gotten audio to work in android with spdif or usb dac? |